A PCR-based protocol for the detection of Leishmania (Viannia) parasites in canine blood, buffy coat, and bone marrow was developed and was then tested with field samples taken from a random sample of 545 dogs from villages in Peru where Leishmania (Viannia) braziliensis and Leishmania (Viannia) peruviana are endemic. Comparative tests with cultured parasites mixed with dog blood showed that the PCR assay's sensitivity was significantly dependent on the DNA extraction protocol and the PCR primers used. Mass screening of field samples by the preferred PCR protocol detected American cutaneous leishmaniasis (ACL) in 44 of 545 (8.1%) dogs; 31 of 402 (7.7%), 20 of 223 (9.0%), and 8 of 46 (17.4%) were PCR positive when whole blood, buffy coat, and bone marrow aspirates, respectively, were tested. The high prevalence of Leishmania in both asymptomatic (7.6%) and symptomatic (18.0%) dogs provides further circumstantial evidence for their suspected role as reservoir hosts of ACL and indicates that hematogenous dissemination of parasites may be a more common pathological phenomenon than has previously been acknowledged. However, unlike for zoonotic visceral leishmaniasis, the comparatively low prevalence of Leishmania (Viannia) in the blood of symptomatic dogs indicates that PCR with blood cannot be the "gold standard" for the (mass) screening of samples in epidemiological studies.